[djbw-nvdimm:libnvdimm-pending 12/16] drivers/dax/mapping.c:402:13: error: redefinition of 'dax_lock_page'

From: kernel test robot
Date: Sat Sep 03 2022 - 00:07:54 EST

Next message: kernel test robot: "Re: [PATCH v2 3/3] PCI/PM: Always disable PTM for all devices during suspend"
Previous message: Shawn Guo: "Re: [PATCH v7 1/2] dt-bindings: arm: fsl: Add MSC SM2S-IMX8PLUS SoM and SM2-MB-EP1 Carrier"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

Hi Dan,

FYI, the error/warning was bisected to this commit, please ignore it if it's irrelevant.

tree: https://git.kernel.org/pub/scm/linux/kernel/git/djbw/nvdimm.git libnvdimm-pending
head: ac9d6b3ba0a537b67bdc8e525308dc371da91e1b
commit: fd2e18c05d353b59283b4cb5508bafecd766b304 [12/16] devdax: Move address_space helpers to the DAX core
config: sh-allmodconfig (https://download.01.org/0day-ci/archive/20220903/202209031146.ZUuNeq2r-lkp@xxxxxxxxx/config)
compiler: sh4-linux-gcc (GCC) 12.1.0
reproduce (this is a W=1 build):
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
chmod +x ~/bin/make.cross
# https://git.kernel.org/pub/scm/linux/kernel/git/djbw/nvdimm.git/commit/?id=fd2e18c05d353b59283b4cb5508bafecd766b304
git remote add djbw-nvdimm https://git.kernel.org/pub/scm/linux/kernel/git/djbw/nvdimm.git
git fetch --no-tags djbw-nvdimm libnvdimm-pending
git checkout fd2e18c05d353b59283b4cb5508bafecd766b304
# save the config file
mkdir build_dir && cp config build_dir/.config
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-12.1.0 make.cross W=1 O=build_dir ARCH=sh SHELL=/bin/bash drivers/

If you fix the issue, kindly add following tag where applicable
Reported-by: kernel test robot <lkp@xxxxxxxxx>

All errors (new ones prefixed by >>):

drivers/dax/mapping.c: In function 'dax_associate_entry':
drivers/dax/mapping.c:345:17: error: implicit declaration of function 'get_dev_pagemap_many'; did you mean 'put_dev_pagemap_many'? [-Werror=implicit-function-declaration]
345 | pgmap = get_dev_pagemap_many(pfn, NULL, PHYS_PFN(size));
| ^~~~~~~~~~~~~~~~~~~~
| put_dev_pagemap_many
drivers/dax/mapping.c:345:15: warning: assignment to 'struct dev_pagemap *' from 'int' makes pointer from integer without a cast [-Wint-conversion]
345 | pgmap = get_dev_pagemap_many(pfn, NULL, PHYS_PFN(size));
| ^
drivers/dax/mapping.c: At top level:
>> drivers/dax/mapping.c:402:13: error: redefinition of 'dax_lock_page'
402 | dax_entry_t dax_lock_page(struct page *page)
| ^~~~~~~~~~~~~
In file included from drivers/dax/mapping.c:11:
include/linux/dax.h:187:27: note: previous definition of 'dax_lock_page' with type 'dax_entry_t(struct page *)' {aka 'long unsigned int(struct page *)'}
187 | static inline dax_entry_t dax_lock_page(struct page *page)
| ^~~~~~~~~~~~~
>> drivers/dax/mapping.c:449:6: error: redefinition of 'dax_unlock_page'
449 | void dax_unlock_page(struct page *page, dax_entry_t cookie)
| ^~~~~~~~~~~~~~~
include/linux/dax.h:194:20: note: previous definition of 'dax_unlock_page' with type 'void(struct page *, dax_entry_t)' {aka 'void(struct page *, long unsigned int)'}
194 | static inline void dax_unlock_page(struct page *page, dax_entry_t cookie)
| ^~~~~~~~~~~~~~~
>> drivers/dax/mapping.c:469:13: error: redefinition of 'dax_lock_mapping_entry'
469 | dax_entry_t dax_lock_mapping_entry(struct address_space *mapping, pgoff_t index,
| ^~~~~~~~~~~~~~~~~~~~~~
include/linux/dax.h:198:27: note: previous definition of 'dax_lock_mapping_entry' with type 'dax_entry_t(struct address_space *, long unsigned int, struct page **)' {aka 'long unsigned int(struct address_space *, long unsigned int, struct page **)'}
198 | static inline dax_entry_t dax_lock_mapping_entry(struct address_space *mapping,
| ^~~~~~~~~~~~~~~~~~~~~~
>> drivers/dax/mapping.c:512:6: error: redefinition of 'dax_unlock_mapping_entry'
512 | void dax_unlock_mapping_entry(struct address_space *mapping, pgoff_t index,
| ^~~~~~~~~~~~~~~~~~~~~~~~
include/linux/dax.h:204:20: note: previous definition of 'dax_unlock_mapping_entry' with type 'void(struct address_space *, long unsigned int, dax_entry_t)' {aka 'void(struct address_space *, long unsigned int, long unsigned int)'}
204 | static inline void dax_unlock_mapping_entry(struct address_space *mapping,
| ^~~~~~~~~~~~~~~~~~~~~~~~
>> drivers/dax/mapping.c:656:14: error: conflicting types for 'dax_layout_pinned_page_range'; have 'struct page *(struct address_space *, loff_t, loff_t)' {aka 'struct page *(struct address_space *, long long int, long long int)'}
656 | struct page *dax_layout_pinned_page_range(struct address_space *mapping,
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
include/linux/dax.h:175:1: note: previous definition of 'dax_layout_pinned_page_range' with type 'struct page *(struct address_space *, long unsigned int, long unsigned int)'
175 | dax_layout_pinned_page_range(struct address_space *mapping, pgoff_t start,
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
>> drivers/dax/mapping.c:718:14: error: redefinition of 'dax_layout_pinned_page'
718 | struct page *dax_layout_pinned_page(struct address_space *mapping)
| ^~~~~~~~~~~~~~~~~~~~~~
include/linux/dax.h:169:28: note: previous definition of 'dax_layout_pinned_page' with type 'struct page *(struct address_space *)'
169 | static inline struct page *dax_layout_pinned_page(struct address_space *mapping)
| ^~~~~~~~~~~~~~~~~~~~~~
drivers/dax/mapping.c: In function 'dax_writeback_one':
>> drivers/dax/mapping.c:904:17: error: implicit declaration of function 'pfn_mkclean_range' [-Werror=implicit-function-declaration]
904 | pfn_mkclean_range(pfn, count, index, vma);
| ^~~~~~~~~~~~~~~~~
cc1: some warnings being treated as errors

vim +/dax_lock_page +402 drivers/dax/mapping.c

324
325 /*
326 * When it is called in dax_insert_entry(), the cow flag will indicate that
327 * whether this entry is shared by multiple files. If so, set the page->mapping
328 * FS_DAX_MAPPING_COW, and use page->index as refcount.
329 */
330 static vm_fault_t dax_associate_entry(void *entry,
331 struct address_space *mapping,
332 struct vm_fault *vmf, unsigned long flags)
333 {
334 unsigned long size = dax_entry_size(entry), pfn, index;
335 struct dev_pagemap *pgmap;
336 int i = 0;
337
338 if (IS_ENABLED(CONFIG_FS_DAX_LIMITED))
339 return 0;
340
341 if (!size)
342 return 0;
343
344 pfn = dax_to_pfn(entry);
> 345 pgmap = get_dev_pagemap_many(pfn, NULL, PHYS_PFN(size));
346 if (!pgmap)
347 return VM_FAULT_SIGBUS;
348
349 index = linear_page_index(vmf->vma, ALIGN(vmf->address, size));
350 for_each_mapped_pfn(entry, pfn) {
351 struct page *page = pfn_to_page(pfn);
352
353 if (flags & DAX_COW) {
354 dax_mapping_set_cow(page);
355 } else {
356 WARN_ON_ONCE(page->mapping);
357 page->mapping = mapping;
358 page->index = index + i++;
359 }
360 }
361
362 return 0;
363 }
364
365 static void dax_disassociate_entry(void *entry, struct address_space *mapping,
366 bool trunc)
367 {
368 unsigned long size = dax_entry_size(entry), pfn;
369 struct page *page;
370
371 if (IS_ENABLED(CONFIG_FS_DAX_LIMITED))
372 return;
373
374 if (!size)
375 return;
376
377 page = pfn_to_page(dax_to_pfn(entry));
378 put_dev_pagemap_many(page->pgmap, PHYS_PFN(size));
379
380 for_each_mapped_pfn(entry, pfn) {
381 page = pfn_to_page(pfn);
382 WARN_ON_ONCE(trunc && page_maybe_dma_pinned(page));
383 if (dax_mapping_is_cow(page->mapping)) {
384 /* keep the CoW flag if this page is still shared */
385 if (page->index-- > 0)
386 continue;
387 } else
388 WARN_ON_ONCE(page->mapping && page->mapping != mapping);
389 page->mapping = NULL;
390 page->index = 0;
391 }
392 }
393
394 /*
395 * dax_lock_page - Lock the DAX entry corresponding to a page
396 * @page: The page whose entry we want to lock
397 *
398 * Context: Process context.
399 * Return: A cookie to pass to dax_unlock_page() or 0 if the entry could
400 * not be locked.
401 */
> 402 dax_entry_t dax_lock_page(struct page *page)
403 {
404 XA_STATE(xas, NULL, 0);
405 void *entry;
406
407 /* Ensure page->mapping isn't freed while we look at it */
408 rcu_read_lock();
409 for (;;) {
410 struct address_space *mapping = READ_ONCE(page->mapping);
411
412 entry = NULL;
413 if (!mapping || !dax_mapping(mapping))
414 break;
415
416 /*
417 * In the device-dax case there's no need to lock, a
418 * struct dev_pagemap pin is sufficient to keep the
419 * inode alive, and we assume we have dev_pagemap pin
420 * otherwise we would not have a valid pfn_to_page()
421 * translation.
422 */
423 entry = (void *)~0UL;
424 if (S_ISCHR(mapping->host->i_mode))
425 break;
426
427 xas.xa = &mapping->i_pages;
428 xas_lock_irq(&xas);
429 if (mapping != page->mapping) {
430 xas_unlock_irq(&xas);
431 continue;
432 }
433 xas_set(&xas, page->index);
434 entry = xas_load(&xas);
435 if (dax_is_locked(entry)) {
436 rcu_read_unlock();
437 wait_entry_unlocked(&xas, entry);
438 rcu_read_lock();
439 continue;
440 }
441 dax_lock_entry(&xas, entry);
442 xas_unlock_irq(&xas);
443 break;
444 }
445 rcu_read_unlock();
446 return (dax_entry_t)entry;
447 }
448
> 449 void dax_unlock_page(struct page *page, dax_entry_t cookie)
450 {
451 struct address_space *mapping = page->mapping;
452 XA_STATE(xas, &mapping->i_pages, page->index);
453
454 if (S_ISCHR(mapping->host->i_mode))
455 return;
456
457 dax_unlock_entry(&xas, (void *)cookie);
458 }
459
460 /*
461 * dax_lock_mapping_entry - Lock the DAX entry corresponding to a mapping
462 * @mapping: the file's mapping whose entry we want to lock
463 * @index: the offset within this file
464 * @page: output the dax page corresponding to this dax entry
465 *
466 * Return: A cookie to pass to dax_unlock_mapping_entry() or 0 if the entry
467 * could not be locked.
468 */
> 469 dax_entry_t dax_lock_mapping_entry(struct address_space *mapping, pgoff_t index,
470 struct page **page)
471 {
472 XA_STATE(xas, NULL, 0);
473 void *entry;
474
475 rcu_read_lock();
476 for (;;) {
477 entry = NULL;
478 if (!dax_mapping(mapping))
479 break;
480
481 xas.xa = &mapping->i_pages;
482 xas_lock_irq(&xas);
483 xas_set(&xas, index);
484 entry = xas_load(&xas);
485 if (dax_is_locked(entry)) {
486 rcu_read_unlock();
487 wait_entry_unlocked(&xas, entry);
488 rcu_read_lock();
489 continue;
490 }
491 if (!entry || dax_is_zero_entry(entry) ||
492 dax_is_empty_entry(entry)) {
493 /*
494 * Because we are looking for entry from file's mapping
495 * and index, so the entry may not be inserted for now,
496 * or even a zero/empty entry. We don't think this is
497 * an error case. So, return a special value and do
498 * not output @page.
499 */
500 entry = (void *)~0UL;
501 } else {
502 *page = pfn_to_page(dax_to_pfn(entry));
503 dax_lock_entry(&xas, entry);
504 }
505 xas_unlock_irq(&xas);
506 break;
507 }
508 rcu_read_unlock();
509 return (dax_entry_t)entry;
510 }
511
> 512 void dax_unlock_mapping_entry(struct address_space *mapping, pgoff_t index,
513 dax_entry_t cookie)
514 {
515 XA_STATE(xas, &mapping->i_pages, index);
516
517 if (cookie == ~0UL)
518 return;
519
520 dax_unlock_entry(&xas, (void *)cookie);
521 }
522

--
0-DAY CI Kernel Test Service
https://01.org/lkp

Next message: kernel test robot: "Re: [PATCH v2 3/3] PCI/PM: Always disable PTM for all devices during suspend"
Previous message: Shawn Guo: "Re: [PATCH v7 1/2] dt-bindings: arm: fsl: Add MSC SM2S-IMX8PLUS SoM and SM2-MB-EP1 Carrier"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]